My Skills

Needs elicitation

Finding product market fit comes from understanding user needs and the difference between those and apparent wants. I like to work with clients to truly understand their processes and better understand where they need help with their data.

Persona Creation

There's a real benefit to getting to know the end users of a data product. I have lots of experience of getting to know clients, building rapport and getting a good understanding of their pain points. I can then leverage this understanding to improve the impact of the data work.

Roadmapping

My recent work has required balancing workloads and working with varied teams to progress multi-disciplinary projects through to their next stage. By understanding impact and development complexity I can better judge the next best step to take.

Stakeholder management

My previous roles have given me experience in communicating to a wide range of stakeholders across the management strata. I have given many presentations to C-suite managers as well as talked to whole teams, stakeholder groups and I have been on trade missions on behalf of the UK.

Protyping

Taking an idea from the inception/beer mat stage through testing and initial deployment into the real world. This includes presenting to potential clients in smoke-test MVP testing.

Deployment

One of my greatest skills, I think, is getting a new product out into the real world. It takes a different level of skills to take a locally tested product and put it on a platform where it can be accessed.

Key Metric Creation

I have lots of experience in examining data and producing key metrics that speak to the underlying needs of an organisation.

Telling the data's story

The nuance of the data will tell a story. Through data analysis I can understand the flow and the story of the data and then present the conclusions.

Data Quality

I can bridge the gap between the data produced and the needs of the end users to help assess if the data is fit for purpose and what might need to be adjusted to make it fit the needs.

Juptyer Lab/Notebook

Using in the cloud or setting up local envs for testing. Jupyter has become my go-to for examining data if the data is not already in a database.

Excel

I have a lot of experience in using excel. Mostly I use jupyter or other environments to examine data these days but I still rank my excel skills as being very high.

VBA

I have created lots of functions in VBA to handle data communications between excel and databases.

SQL

I have a lot of experience using SQL to examine data and to create reports. This includes using LookerML, MySQL, Postgres, SQLite, MSSQL.

Understanding the audience

For me, the most important skill is putting yourself in the shoes of the audience. You can maximise the impact of the data visualisation by envisaging the persons and understanding how they will react. I also like to look to the "nudge points" of the audience, i.e understanding what it is I want them to do when they see the visualisation.

The Killer Chart

When presenting data to an audience you often only get a small window to present the overall findings of your research. I enjoy sweating the data to get to the "killer chart", the one that folks will remember and that offers the actionable insight.

Videos

Nothing holds the attention like a video of information. I like presenting data using videos so that you can tell the story of data in four dimensions, i.e. show how the data's story evolves over time.

Infographics

Focussing on the story and how users consume that data is important. Infographics offer a great way to take people on a journey, or a way to demonstrate the context of the data.

Powerpoint

The one we love to hate. From my work at an investment bank I know the importance of a strong slidedeck that has the right data presented along with a strong narrative and tempo.

D3.js

D3 is a great way of visualisation data in a web environment. Ok, it's verbose and has a lot of functionality but once you've learnt the basics it is a powerful tool for visualisation and SVG manipultation.

Matplotlib

The go-to visualisation for doing exploratory data analysis and more complex data reporting requirements. I spend a lot of time outputting charts into PDF reports and matplotlib has great functionality for this.

Seaborn

Not just a cool overlay for Matplotlib, I often use the statistic representation functions to conduct quick drill downs into the data. The correlation plots are awesome for getting a better understanding of features that could be used for model training.

Looker

I've used Looker a lot but I've still got more to learn. I like a lot of the functionality of Looker. I feel that it's a tool that gets overused and can often hide data errors in views. It is a good tool that I like using for standardised data visualisation and reporting.

Bokeh

This is a cool library that I spent a fair amount of time trialling out. You get a lot of functionality for drilling into the data in the visualisations. I often go straight to D3 if I need more user input to get into the data.

Azure

Most of my experience has been using the Azure products to conduct data science experiments and train ML models. I think that their offering in this area is my current favourite in terms of speed to get to a useable model and track experiment progression.

GCP

I have used GCP mostly for BigQuery work. I've experimented a lot with the data products and visualisation offering. I think it's my preferred cloud offering based on its user experience and interface.

I am very excited aboud BigQueryML and look forward to using SQL based ML in the future.

AWS

I am a relative newcomer to AWS but have spent nearly a solid month just using it so that I could better understand what was on offer.

I have nearly completed their DevOps course on Coursera (available here) which has taught me a lot about the products they have on offer.

IBM Cloud

My first cloud. They don't hold market-share dominance right now but I do really like their redshift offering.

Scikit Learn

My go-to machine learning package. Who doesn't love a good decision tree? I really like breaking business process problems into series of ML tasks. In my view the simplist model is the most powerful.

Auth0

This is my current favourite option for securing apps. The SDK is an easy route to bring industry leading Identity Access Management into an app.

Python

Writing code in python has been a great experience. I have learned a lot and grown in confidence over the last 6 years that I have been coding.

JavaScript

I learnt JavaScript to enable me to produce apps using electron.js. Since then I have achieved the same level of skill in JS as I have in python. Particularly when it comes to data visualistion through other packages like D3.js. Building and testing APIs (with Express of with web functions) is at the heart of my recent work in putting data insights in the place where they need to be.

Node.JS

I love node and node package manager. The speed at which you can develop ideas into apps is amazing. I am trying to transition towards TDD (Test Driven Development) to get to production ready code earlier and reduce debugging time.

C#

I have learnt C# so that I can write functions for my work in Unity for game development.

CSS

The world would be pretty bland without CSS. I've been trying bootstrap (which I've used for this site) to make coding CSS easier but I still find myself falling back on "coding" presentation using raw CSS, particularly when it comes to adjusting visualisations in D3.js etc.

HTML

Seems a bit obvious to say but no harm in listing it.

Electron.js

I think electron.js is really cool. the ability to use the same (or similar stack) as you would for a web app to build a native desktop app is a really powerful tool. I've build a number of desktop apps and I think it's got a big future in tailoring data product solutions to meet on-prem requirements.

Flask

For building python-based APIs this is my favourite solution. I find it particularly powerful for getting scikit-learn models into the real world.

Github

I took one of the coursera courses on Github offered by Google (available here). I thought that I knew a fair bit before hand but that course showed me some of the real power of Git and the remote repos.

Github is my preferred repo but I've also tried CodeCommit from AWS.

Jest

My transition to a more mature programmer has meant that I think about security and testing before I start developing code. I use Jest for testing my node apps.

Mongo DB

Being able to store data in the same format that you use it in a web-first world is so powerful. I really like the different libraries available for Mongo, like PyMongo and Mongoose.

Pytest

I haven't used pytest as much as I ought to. For an AWS ML model in lambda function project I've been able to greatly improve my skills with pytest.

Unity

This could be my new favourite thing. I think that the future of data products will start to use more novel forms of representation such as game-based interactions.

My recent work in Unity has been for a project to capture information within a game and transport it to a MongoDB backend throug an Auth0 secured app.